How umeshchandra-rao/scanlance works
Project Type & Purpose
• Full-stack AI-driven resume analysis web app.
• Frontend built with Next.js 14 App Router, React, Tailwind CSS, Framer Motion, shadcn/ui.
• Backend serverless API routes (Next.js) orchestrate file extraction, AI calls (Google Gemini), rate-limiting (Redis), and feedback submission.
High-Level Components
a. Frontend UI Layer
• App Router (app/layout.tsx, app/page.tsx, app/analyzer/page.tsx)
• Analyzer feature (app/analyzer/…): pages, layouts, animations, constants, utils
• Shared components (components/… & components/ui/…): hero, navbar, footer, file uploader, interactive result views
• Custom hooks (app/analyzer/hooks, hooks/…): useAnalysis, useFileHandling, useStepNavigation, useToast, useMobile
b. Backend API Layer
• API routes (app/api/analyze/route.ts, app/api/send-feedback/route.ts)
• Server actions (app/actions.ts) for form handling
c. Business Logic & Services
• lib/: analyze-resume.ts, gemini-client.ts, pdf-extractor.ts, docx-extractor.ts, text-extractor-utils.ts
• Domain-specific NER modules: skill-taxonomy.ts, domain-specific-ner.ts, education-analyzer.ts, improved-resume-section-detector.ts
• Rate limiter (lib/rate-limiter.ts) with optional Redis
• Validation & error handling (validation-schemas.ts, error-handler.ts, env-validation.ts)
d. External Services
• Google Gemini API for AI analysis
• Redis (optional) for rate limiting
• Resend API (optional) for feedback emails
e. Configuration & Assets
• next.config.mjs, tailwind.config.ts, tsconfig.json, .env(.example)
• Public/static assets (public/) and global styles (styles/globals.css)
Component Relationships & Data Flows
• User interacts with Frontend UI → uploads resume + enters job description
• Frontend invokes client-side hooks (useFileHandling, useAnalysis) → submits POST /api/analyze
• Next.js API route (/api/analyze)
→ validates input (Zod schemas)
→ invokes lib.analyzeResume:
• runs extractors (PDF, DOCX, TXT)
• runs NER and domain analyzers
• constructs AI prompts (prompt-templates.ts)
→ calls Google Gemini client → receives AI response
→ applies scoring & recommendations → returns JSON
• Frontend receives JSON → ResultsDisplay component renders interactive breakdown
• Feedback loop: user submits feedback → POST /api/send-feedback → optional Resend email dispatch
Architectural Patterns & Principles
• Next.js App Router for file-based routing & nested layouts
• Serverless API endpoints co-located with pages
• Separation of concerns: UI components vs. domain/business logic (lib/) vs. data-flow hooks
• Domain-driven modules in lib/ for extraction, analysis, validation
• Progressive enhancement: optional Redis integration with graceful fallback
• Strict typing & schema validation (TypeScript + Zod)